Prospective Console Recruitment

[Exophase]

Hi everyone,

I've been talking with someone who is planning and funding a new handheld gaming platform. In order for this to get off the ground he is going to need some electrical and mechanical engineers for development and design work. So I figured that perhaps, in the same way Craig found his excellent electrical and mechanical design talent for the Pandora on these boards (MWeston and DaveC respectively) I might have some luck recruiting here on his behalf.

Forewarning: I am in fact posting this strictly in the interest of getting help from a small number of highly skilled individuals; it is not my intention to begin promoting/hyping this console. It's only in the early brainstorming stages, so take all of this accordingly.

That said, I was asked to specify a bit of what's planned so far in order to give a better idea of the general goals and design philosophy and help determine whether or not this is something that is of interest to anyone. So this is what's currently on the table:

eon gaming handheld
CPU: i.MX538 @ 1GHz (Freescale specifies "up to 1.2GHz", this is kind of a wait and see), including hardware 1080p decode
RAM: 512MB of 400MHz LPDDR2, DDR3, or DDR2 (in that order of preference from most to least desired; trying to find what's any kind of option in terms of pricing)
Display: 4.8" 800x480 w/resistive touch (current candidate: CPT CLAA048LA0BCT)
Controls: dpad, 4 face buttons, front + back shoulders, 2x miniature analog sticks (not nubs) using CTS 254 joystick potentiometers
Motion: Triaxial accelerometer, Freescale MMA8450 or MMA8452 (depending on what makes sense in terms of voltage levels)
Control interface: STM32F101 Cortex-M3
PMIC: Holding out for the not-yet-production MC34708 (might settle for Dialog DA9053 or something from Linear Tech)
Audio: Freescale SGTL5000 codec (or possibly something better)
Wifi/BT: Atheros AR6102
Internal storage: 2GB micro-SD
External storage: 1x SD
TV-out: HDMI, with VGA being a potential option over ext connector (i.MX53 has support for it on-chip)
Connectivity: USB OTG, USB host external connectors, USB host internal connector inside battery compartment
Battery: 2600+mAh Li-Po/Li-Ion
OS: Ubuntu and/or Android (and/or Angstrom? - note the first two have BSPs from Freescale already)
Price: Target of $200 in production (initial batch of ~500 may cost more, ~$250+?)

And it is my personal wish that if this does get off the ground that it attempts to cross promote software development on the Pandora, since it'll hopefully match decently well for things that don't rely on keyboard.

For those interested, here is a preliminary concept sketch:

http://imageshack.us...844/protok.png/

It's really just tentative and an idea of basic shape, nothing strictly designing to w/o input of mechanical design and board dimensions. Note that this design has a heavier emphasis on comfort than compact size; the general feeling is that there's plenty of the latter on the market and not a lot of the former.

So, if this appeals to anyone skilled in EE or mechanical CAD design please either contact me or find us on #eon at irc.synirc.net.

 

[Nintendo]

Spoiler

Hi everyone,

I've been talking with someone who is planning and funding a new handheld gaming platform. In order for this to get off the ground he is going to need some electrical and mechanical engineers for development and design work. So I figured that perhaps, in the same way Craig found his excellent electrical and mechanical design talent for the Pandora on these boards (MWeston and DaveC respectively) I might have some luck recruiting here on his behalf.

Forewarning: I am in fact posting this strictly in the interest of getting help from a small number of highly skilled individuals; it is not my intention to begin promoting/hyping this console. It's only in the early brainstorming stages, so take all of this accordingly.

That said, I was asked to specify a bit of what's planned so far in order to give a better idea of the general goals and design philosophy and help determine whether or not this is something that is of interest to anyone. So this is what's currently on the table:

eon gaming handheld
CPU: i.MX538 @ 1GHz (Freescale specifies "up to 1.2GHz", this is kind of a wait and see), including hardware 1080p decode
RAM: 512MB of 400MHz LPDDR2, DDR3, or DDR2 (in that order of preference from most to least desired; trying to find what's any kind of option in terms of pricing)
Display: 4.8" 800x480 w/resistive touch (current candidate: CPT CLAA048LA0BCT)
Controls: dpad, 4 face buttons, front + back shoulders, 2x miniature analog sticks (not nubs) using CTS 254 joystick potentiometers
Motion: Triaxial accelerometer, Freescale MMA8450 or MMA8452 (depending on what makes sense in terms of voltage levels)
Control interface: STM32F101 Cortex-M3
PMIC: Holding out for the not-yet-production MC34708 (might settle for Dialog DA9053 or something from Linear Tech)
Audio: Freescale SGTL5000 codec (or possibly something better)
Wifi/BT: Atheros AR6102
Internal storage: 2GB micro-SD
External storage: 1x SD
TV-out: HDMI, with VGA being a potential option over ext connector (i.MX53 has support for it on-chip)
Connectivity: USB OTG, USB host external connectors, USB host internal connector inside battery compartment
Battery: 2600+mAh Li-Po/Li-Ion
OS: Ubuntu and/or Android (and/or Angstrom? - note the first two have BSPs from Freescale already)
Price: Target of $200 in production (initial batch of ~500 may cost more, ~$250+?)

And it is my personal wish that if this does get off the ground that it attempts to cross promote software development on the Pandora, since it'll hopefully match decently well for things that don't rely on keyboard.

For those interested, here is a preliminary concept sketch:

http://imageshack.us...844/protok.png/

It's really just tentative and an idea of basic shape, nothing strictly designing to w/o input of mechanical design and board dimensions. Note that this design has a heavier emphasis on comfort than compact size; the general feeling is that there's plenty of the latter on the market and not a lot of the former.

So, if this appeals to anyone skilled in EE or mechanical CAD design please either contact me or find us on #eon at irc.synirc.net.

HOLY MOLY!!! Although I don't agree with all the specs, etc...It still sounds HOT AND SEXY!!! :wub:
If and when it passes the brainstorming stage, would I be able to donate some money to this project?

 

[Exophase]

The guy who is heading it has a fair bit of cash behind it, so I think it's okay to get it off the ground. Probably not the sort of thing that'll be asking for open donations.

Just out of curiosity, what would you change with the specs? Bearing in mind, while I'm sure he'd love to have a cutting edge dual core Cortex-A9 in there those aren't really available in low volume. In fact, some of the only higher end SoCs we know of are TIs OMAP35xx, AM/DM37xx, and Freescale's i.MX51/53. The first of both set are basically first generation 65nm Cortex-A8, while the second are second generation 45nm with higher CPU/GPU/memory clocks and some more features (and memory and what have you).

If anyone has any other questions I can probably field them, to the best of my knowledge.

(regarding the Pandora boards, see my PM response)

 

[Nintendo]

The guy who is heading it has a fair bit of cash behind it, so I think it's okay to get it off the ground. Probably not the sort of thing that'll be asking for open donations.

Just out of curiosity, what would you change with the specs? Bearing in mind, while I'm sure he'd love to have a cutting edge dual core Cortex-A9 in there those aren't really available in low volume. In fact, some of the only higher end SoCs we know of are TIs OMAP35xx, AM/DM37xx, and Freescale's i.MX51/53. The first of both set are basically first generation 65nm Cortex-A8, while the second are second generation 45nm with higher CPU/GPU/memory clocks and some more features (and memory and what have you).

If anyone has any other questions I can probably field them, to the best of my knowledge.

(regarding the Pandora boards, see my PM response)

1. I've edited something out of my previous post

2. So...he's rich?

3. My changes consist of the following (sorry, me being a high-end specification whore I can't bear in mind any cutting edge limitations - expensive and overkill, I know):

• 1GB RAM
• 5.5" display @ 1024 x 480
• Capacitive touch screen

I also hope for the following:

• AMOLED screen
• SDXC support (all specs)
• Invisible speakers
• 64GB SSD internal storage
• 2 x USB 3.0 ports
• BT 3.0

Questions:

- Does using CTS 254 joystick potentiometers mean you can't use them as buttons like PlayStation's dual analogue sticks (L3 & R3)?
- Can you get super-accurate analgue controls using CTS 254 joystick potentiometers ( http://media.digikey.com/photos/CTS%20Photos/254%20Series.jpg )?
- Triaxial accelerometer with Gyro?
- Would over-clocking to 2Ghz be possible?

 

[Exophase]

Okay, I'll try to address your questions.

2. So...he's rich?

He sounds like he's done pretty well for himself, with a phone repairs shop he has setup.. he wants to move into manufacturing these sorts of devices instead, so he's pretty intent on pushing what he can into it.

3. My changes consist of the following (expensive and overkill, I know):

• 1GB RAM
• 5.5" display @ 1024 x 480
• Capacitive touch screen

1GB RAM has been considered, but the costs are pretty uncertain right now and it can make a big difference. Another consideration is that if you have to build it from 8bit parts (due to pricing) then building up enough for 1GB can be prohibitive, for instance this thing can't have 8 RAM chips on board.

5.5" is pretty non-standard, but at that point you really start using a lot more power for backlighting and the LCD itself (even 4.8" is pushing it).. 1024x480 is not a good resolution, you'd want 1024x600. But actually, even getting 800x480 below 7" is not as common as you'd think, in easily available commodity parts that aren't cell phone replacement screens. Capacitive touch has been kicked around, but this is also something you don't see much of on commodity screens as an option (it's much better to get touch with the LCD if you can), and you need special controllers for it (touch comes standard on a lot of SoCs or PMICs, or if not can be easily interfaced with a simple lowish resolution 4-channel ADC).. that and the team doesn't seem to really like capacitive anyway.

I also hope for the following:

• AMOLED screen
• SDXC support (all specs)
• Invisible speakers
• 64GB SSD internal storage
• 2 x USB 3.0 ports
• BT 3.0

The only realistic AMOLED option I'm aware of is Samsung's 3.7" 800x480 featured in phones like Nexus One and what have you. And that is pentile, or half the resolution for red and blue pixels. In general, Chinese LCD manufacturers are going to give you the cheapest prices, as opposed to Korean (Samsung, LG) or Japanese (Mitsubishi, NEC, Hitachi, Sharp, etc) or god-forbid American (Optrex). An AMOLED would also require an external controller for this SoC, piling on more costs..

SDXC is Not an option on current SoCs I'm aware of, so would mean at best adding another expensive component for something few would benefit from.

Not really sure what you mean by invisible speakers.

64GB SSD would be very expensive and probably well outside the power budget of the unit. You could expect to see something like this on high end tablets but not handhelds, at least not those that have to dedicate their power to other things.

USB 3.0 is not an SoC option yet, same story as with SDXC.. I'm not sure what the power requirements would be to benefit from the higher bandwidth either, you wouldn't expect that to come for free.

I doubt you'll find BT 3.0 in any reasonably priced wifi + BT module.. and furthermore I'm not sure why you'd really want it.

When you get down to it, all of these cutting edge features are just not possible on an affordable handheld, and moreover aren't even really available to low volume small scale startups. This has to be done with strictly off the shelf parts, preferably ones available from Chinese suppliers (if not actually made in China) as much as possible.

- Does using CTS 254 joystick potentiometers mean you can't use them as buttons like PlayStation's dual analogue sticks (L3 & R3)?

Actually, CTS offers versions with buttons, which I think they'll be using.

- Can you get super-accurate analgue controls using CTS 254 joystick potentiometers ( http://media.digikey...54%20Series.jpg )?

I don't really know what the limitations are, but my guess is that so long as the resistance properties are reasonably stable (or you recalibrate as necessary) it'll mainly be down to the ADC, which in this case will give 4096 positions with decent accuracy, Beyond that I'm sure mechanical design will play a role, in how they're capped. One thing is that the stick control software would be fully slaved by the SoC, meaning that it'll be possible to update it and avoid issues with autocalibration and whatever (IMO, using microcontrollers with fixed firmware on the Pandora's nubs was not good)

- Triaxial accelerometer with Gyro?

Triaxial accelerometer yes, gyros (which are much more expensive), probably not, but I guess it'll depend on pricing and availability.

- Would over-clocking to 2Ghz be possible?

No, absolutely not ;p I doubt you'll ever see a Cortex-A8 go anywhere near that speed in this process node. Freescale says we might see official speeds as high as 1.3GHz, although I really don't know for which parts and what context that applies to. From the power draw numbers I'm sure you'd REALLY kill the battery at that, even 1GHz is probably going to really push it.

 

[notaz]

And it is my personal wish that if this does get off the ground that it attempts to cross promote software development on the Pandora, since it'll hopefully match decently well for things that don't rely on keyboard.

In that case you should consider adopting GPH-style game button layout and Ubuntu would be better choice over Android.

 

[sebt3]

In that case you should consider adopting GPH-style game button layout and Ubuntu would be better choice over Android.

+1
look nice and interesting

 

[Kamillus]

Hi, I think it would be a good alternative to Pandora if the price will be ~200$ with those specs.

- The 4.8" screen is a bit too big unless it could be a slider type handheld, but i would recommend maximum 4.3" with the same resolution. I don't think it will be of much difference in view as with battery consumption.
- When it comes to battery it should be user removable with a high capacity.
- The triaxial accelerometer could be replaced with a g-sensor as I think it has more potential for racing or FPS type games where u can for example target with a console by moving.
- Also I would remove BT and one pair of shoulder buttons and make the remaining pair comfortable to hold.

Either way it looks promising

Sorry for my English.

 

[Exophase]

Hi, I think it would be a good alternative to Pandora if the price will be ~200$ with those specs.

- The 4.8" screen is a bit too big unless it could be a slider type handheld, but i would recommend maximum 4.3" with the same resolution. I don't think it will be of much difference in view as with battery consumption.
- When it comes to battery it should be user removable with a high capacity.
- The triaxial accelerometer could be replaced with a g-sensor as I think it has more potential for racing or FPS type games where u can for example target with a console by moving.
- Also I would remove BT and one pair of shoulder buttons and make the remaining pair comfortable to hold.

Either way it looks promising

Sorry for my English.

Hi Kamillus,

4.3" vs 4.8" is currently going to come down more to what LCDs are available; there's a careful balance of price and specs behind the decision first and foremost. One thing you'll find with the slightly larger ones is that they tend to have better response time at somewhat of a cost to power consumption, and they tend to have higher maximum brightness (at a more substantial cost to power consumption, but obviously you can dial this down). One of the goals is to try to get at least a little better response time than Pandora's 4.3" TD043MTEA1 LCD, in order to reduce ghosting. For comparison, Pandora's LCD has a typical response time of 10 + 30ms, while this one has a typical of 20ms (although it doesn't specify how it's divided between rise and fall; the closer it is to balanced the better, so if it's 10 + 10 instead of 5 + 15 it'll be more than twice better than the TD043MTEA1's. Of course, this is just going on datasheet listing, they'd want to try things for themselves.

For battery they're going for something that's higher capacity than say, typical phones, but they'd also like something that's already available (not custom made) in order to improve availability down the road/reduce price. Since this handheld will be longer than Pandora I imagine the battery can afford to be too, but at the same time they want to avoid making it the same thickness as Pandora's base, which its battery adds a lot to. In my opinion, 10+ hours of battery life is great but overkill and you can make due with less, but they'll probably have to determine what makes sense based on what the power consumption is really like.

From what I understand "g sensor" is just another name for accelerometer, only often at less than three axes. It isn't worth recouping a very small cost (if any really) for sacrificing degrees of freedom. A triaxial accelerometer should still do OK for motion controls if it's sensitive enough, although gyroscopes are better; maybe that's what you're asking for. I know back 2 years ago when we used a gyro on a product at work all that was available were Invensense products that were pretty expensive, and at that time you couldn't get three-axis either, but these days there are at least gyros from ST and Analog Devices. But even the cheapest ones (which appear to be from ST) still cost 4-5 times an accelerometer does and are analog output so you need to attach them to ADCs and have some passives for signal conditioning (digital ones are available but more expensive)

BT is another relatively inexpensive addition since a lot of wifi modules are available in versions with or without it, and going with doesn't cost an awful lot more. Having four shoulder buttons is an important design goal for them, but if back to back doesn't make sense I'm sure they'll transition to side to side instead.

 

[redlemon]

Looks like it could be a really nice device! One thing I would suggest would be to have some kind of larger button on the side of the dpad for vertical play. Similar to the I and II buttons on the Caanoo. I don't know if that could fit along with a dpad and analog stick though. I think their is a market for the size your aiming for. When I take any of my handhelds out I generally carry them in my messenger bag.

 

[Exophase]

Looks like it could be a really nice device! One thing I would suggest would be to have some kind of larger button on the side of the dpad for vertical play. Similar to the I and II buttons on the Caanoo. I don't know if that could fit along with a dpad and analog stick though. I think their is a market for the size your aiming for. When I take any of my handhelds out I generally carry them in my messenger bag.

Maybe that select-like button could be repositioned, then the analog with a stick click would act as another button. I'll see what they think about this. I agree that vertical play on an emulator-focused handheld is a nice feature.

 

[redlemon]

Other than that I think the design is fairly solid. I like how the speaker holes seem to curve from the bottom upwards. It looks like a nice compromise to the usually necessary awkward positioning of speakers on handheld devices.

 

[DisgruntleElf]

Will the two sticks be able to switch to digital as well? If this device is all about emulation and retro gaming(MAME in particular) then solid digital sticks would be a better choice. Analog sticks and nubs will eventually drift after extensive use.

 

[Exophase]

Will the two sticks be able to switch to digital as well? If this device is all about emulation and retro gaming(MAME in particular) then solid digital sticks would be a better choice. Analog sticks and nubs will eventually drift after extensive use.

They're potentiometers, so electrically the won't appear digital. I presume drift would occur due to a bias shift in the potentiometer resistances due to aging, which would be fixed via recalibration. I don't actually see the difference between a digital stick and an analog stick with extremely tight deadzone and low resolution. Since the stick firmware would be modifiable and probably open source I imagine it'll be possible to artificially degrade/confine the stick at a low level that affected all software (or turn it off entirely in the event that they flat out broke)

 

[fusion_power]

Very interesting.
At first I thought you are talking about Bobs strange "ND" Project but this here seems to make much more sense. I can't judge the device just from it's specs or from the early design Sketch. Looks like a mix of an PSP and an classic megadrive Controller. Maybe some more advanced 3D Renderings can make this more clear but I like the Megadrive Gamepad so it is not the worst Idea. ^^ It also looks overall more rounded than the more ore less "edgy" Pandora. AND this Handheld seems to be able to be hold verticaly too, like GP2X, Caanoo... big advantage over Pandora when it comes to vertical arcade shooters.

Hardware specs sound beefy enough for me, I just wish, Pixel QI would make LCDs in this size, would be the perfect thing IMHO ( http://www.pixelqi.com ). Don't use OLEDs.
I don't know the skills of the 3D Unit but I recommend to use a Chip where actualy useful Drivers exist for, and I mean GOOD Drivers. I can't remember when the Pandora ever got an SGX Driver Update, at least nothing that would boost the Software like I know it from some PC GFX Driver Updates. I have good experiences with NVidia I have to say, so maybe Tegra could be an Option?
Same for the WiFi Module. We know from the Pandora that the Chip itself can be as good in theory be but without good drivers you don't get out of what is in. ^^"

However, I wish you all luck with this Project, I guess you can make usage of the experiences from the Pandora history and avoid some mistakes before things get messy.

 

[Exophase]

This project actually came about from people talking on the #nD IRC channel, as something that made more sense/was more realistic. I agree that there's not a lot to pass judgment on it - I'd like to reiterate that I posted this with the naive hope of finding people to help, not to get feedback. Let me know if anyone is actually interested

Tegra 2 isn't much of an option - they require minimum orders of 100,000, it just isn't something you can dive into without a huge budget. This kind of device would really struggle to sell 100,000 and no one backing it will have that kind of money. Tegra's GPU is IMO a bit overrated anyway, and its CPUs lack NEON which is unattractive to me. Exynos 4210 or OMAP4 would be a nicer solution on paper, but I don't know what costs are like for those either.. they're both selling in low volume dev platforms (PandaBoard and ODROID-A respectively), but I imagine those are special arrangements with the companies. Eventually there might be a more widely/low volume available version of the OMAP4, like there was for OMAP3.. but I haven't heard anything.

The 3D drivers are available and work (OpenGL ES 2.0 of course); for reference, the GPU is basically the same hardware as the Adreno 200. But that was usually clocked at 125MHz in Snapdragons, while this is 200MHz. It's a pretty dated arch now, but I think it'll typically perform at least as well as Pandora's also dated 110MHz SGX 530, meaning that it should at least meet the usual emulator targets (N64 really..). Pandora gets almost no 3D software so it's hard to consider this worth prioritizing, otherwise maybe they should have gone with DM3730 instead.

For the wifi module, the plan is to make sure the module is breadboarded very early on an SDIO card or USB dongle if possible, to ensure that the driver support is there and the quality is acceptable.

 

[MWeston]

Looks quite interesting. So what does the job pay?

 

[Exophase]

Looks quite interesting. So what does the job pay?

Heh, I didn't think you'd be interested and didn't think I should ask you.. but if you want to talk to the guy in charge/funder you should check out the IRC channel. Or I can mediate this.

 

[Nintendo]

Exo, I have been reading the thread with great interest and wanted to make a huge post, but I'm still at awe about all this. I have so much to say (a few more questions and some responses to people in this thread), but I don't know where to start.
The potential for all this is just too exhilaration to ignore.

 

[MWeston]

Well, I wouldn't screw over Craig and ED but new projects always put a gleam in my eye. I just love designing electronics. As usual, who knows if these folks really understand the difficulty if they are just researching and looking for people. It's been a long and scary ride that would take a book to explain.

This looks like more of a GPH style device than a Pandora one which is kinda nice.